Anomaly Prediction Based on Machine Learning for Memory-Constrained Devices

Yuto KITAGAWA; Tasuku ISHIGOOKA; Takuya AZUMI

doi:10.1587/transinf.2018EDP7339

Anomaly Prediction Based on Machine Learning for Memory-Constrained Devices

Yuto KITAGAWA, Tasuku ISHIGOOKA, Takuya AZUMI

Full Text Views

0

Cite this

Summary :

This paper proposes an anomaly prediction method based on k-means clustering that assumes embedded devices with memory constraints. With this method, by checking control system behavior in detail using k-means clustering, it is possible to predict anomalies. However, continuing clustering is difficult because data accumulate in memory similar to existing k-means clustering method, which is problematic for embedded devices with low memory capacity. Therefore, we also propose k-means clustering to continue clustering for infinite stream data. The proposed k-means clustering method is based on online k-means clustering of sequential processing. The proposed k-means clustering method only stores data required for anomaly prediction and releases other data from memory. Due to these characteristics, the proposed k-means clustering realizes that anomaly prediction is performed by reducing memory consumption. Experiments were performed with actual data of control system for anomaly prediction. Experimental results show that the proposed anomaly prediction method can predict anomaly, and the proposed k-means clustering can predict anomalies similar to standard k-means clustering while reducing memory consumption. Moreover, the proposed k-means clustering demonstrates better results of anomaly prediction than existing online k-means clustering.

Publication: IEICE TRANSACTIONS on Information Vol.E102-D No.9 pp.1797-1807

Publication Date: 2019/09/01

Publicized: 2019/05/30

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2018EDP7339

Type of Manuscript: PAPER

Category: Artificial Intelligence, Data Mining

Authors

Yuto KITAGAWA
  Osaka University
Tasuku ISHIGOOKA
  Hitachi Ltd.
Takuya AZUMI
  Saitama University

Keyword

anomaly prediction, k-means clustering, online k-means clustering

Cite this

Copy

Yuto KITAGAWA, Tasuku ISHIGOOKA, Takuya AZUMI, "Anomaly Prediction Based on Machine Learning for Memory-Constrained Devices" in IEICE TRANSACTIONS on Information, vol. E102-D, no. 9, pp. 1797-1807, September 2019, doi: 10.1587/transinf.2018EDP7339.
Abstract: This paper proposes an anomaly prediction method based on k-means clustering that assumes embedded devices with memory constraints. With this method, by checking control system behavior in detail using k-means clustering, it is possible to predict anomalies. However, continuing clustering is difficult because data accumulate in memory similar to existing k-means clustering method, which is problematic for embedded devices with low memory capacity. Therefore, we also propose k-means clustering to continue clustering for infinite stream data. The proposed k-means clustering method is based on online k-means clustering of sequential processing. The proposed k-means clustering method only stores data required for anomaly prediction and releases other data from memory. Due to these characteristics, the proposed k-means clustering realizes that anomaly prediction is performed by reducing memory consumption. Experiments were performed with actual data of control system for anomaly prediction. Experimental results show that the proposed anomaly prediction method can predict anomaly, and the proposed k-means clustering can predict anomalies similar to standard k-means clustering while reducing memory consumption. Moreover, the proposed k-means clustering demonstrates better results of anomaly prediction than existing online k-means clustering.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDP7339/_p

Copy

@ARTICLE{e102-d_9_1797,
author={Yuto KITAGAWA, Tasuku ISHIGOOKA, Takuya AZUMI, },
journal={IEICE TRANSACTIONS on Information},
title={Anomaly Prediction Based on Machine Learning for Memory-Constrained Devices},
year={2019},
volume={E102-D},
number={9},
pages={1797-1807},
abstract={This paper proposes an anomaly prediction method based on k-means clustering that assumes embedded devices with memory constraints. With this method, by checking control system behavior in detail using k-means clustering, it is possible to predict anomalies. However, continuing clustering is difficult because data accumulate in memory similar to existing k-means clustering method, which is problematic for embedded devices with low memory capacity. Therefore, we also propose k-means clustering to continue clustering for infinite stream data. The proposed k-means clustering method is based on online k-means clustering of sequential processing. The proposed k-means clustering method only stores data required for anomaly prediction and releases other data from memory. Due to these characteristics, the proposed k-means clustering realizes that anomaly prediction is performed by reducing memory consumption. Experiments were performed with actual data of control system for anomaly prediction. Experimental results show that the proposed anomaly prediction method can predict anomaly, and the proposed k-means clustering can predict anomalies similar to standard k-means clustering while reducing memory consumption. Moreover, the proposed k-means clustering demonstrates better results of anomaly prediction than existing online k-means clustering.},
keywords={},
doi={10.1587/transinf.2018EDP7339},
ISSN={1745-1361},
month={September},}

Copy

TY - JOUR
TI - Anomaly Prediction Based on Machine Learning for Memory-Constrained Devices
T2 - IEICE TRANSACTIONS on Information
SP - 1797
EP - 1807
AU - Yuto KITAGAWA
AU - Tasuku ISHIGOOKA
AU - Takuya AZUMI
PY - 2019
DO - 10.1587/transinf.2018EDP7339
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2019
AB - This paper proposes an anomaly prediction method based on k-means clustering that assumes embedded devices with memory constraints. With this method, by checking control system behavior in detail using k-means clustering, it is possible to predict anomalies. However, continuing clustering is difficult because data accumulate in memory similar to existing k-means clustering method, which is problematic for embedded devices with low memory capacity. Therefore, we also propose k-means clustering to continue clustering for infinite stream data. The proposed k-means clustering method is based on online k-means clustering of sequential processing. The proposed k-means clustering method only stores data required for anomaly prediction and releases other data from memory. Due to these characteristics, the proposed k-means clustering realizes that anomaly prediction is performed by reducing memory consumption. Experiments were performed with actual data of control system for anomaly prediction. Experimental results show that the proposed anomaly prediction method can predict anomaly, and the proposed k-means clustering can predict anomalies similar to standard k-means clustering while reducing memory consumption. Moreover, the proposed k-means clustering demonstrates better results of anomaly prediction than existing online k-means clustering.
ER -